OK, so I have a signal routed from the analog inputs (H/W In 1/2) to the analog outputs (H/W Out 1/2) on the Patchbay/Router tab of the control panel app. The impression is that the signal is being routed directly.

But it isn't so... changing the "Codec Sample Rate" on the "Hardware Settings" tab of the control panel app alters the sound! Clearly, the analog inputs are being sampled and digitized A->D before routing to the analog outputs D->A. This is disappointing, as I could have sworn otherwise. There appears to be no way to listen to the analog signal directly though the card without A->D->A conversion.

Now, I'm going to have to change the "Codec Sample Rate" to 96KHz for potentially best sound when just listening to analog sources, and remember to change it back to 44.1KHz for most other sources (to prevent a 'pop' upon sample rate switching) -- very annoying. Maybe there's no audible difference, but it "feels better" to be listening at 96KHz, of course...

Maybe I'll buy an "A/B" switchbox for listening to analog unaltered, just to be anal about it .

Edit -- anyone know if I would be better off using 88.2 or lower with the 24/96, for any known reason? Or is it as clean at 96KHz as it is at 44.1KHz?

Thank you for the info.

You know that 44.1 / 48 / 96 kHz is a highly debated question, and that no one will give you a definitive answer. At least between 48 and 96.
Pro 96 : the filter is rejected far outside audible range.
96 legend : the jitter is twice inferior. Wrong! It is the same that at 48 kHz.
Con 96 : The high frequency inaudible noise can intermodulate and create audible distiortion that you'll avoid at 48 kHz.

The only thing merely sensible is to use 48 kHz 24 bits instead of 44.1 kHz 16 bits.

Any card that does that?

Everything is mixed digitally before the DACs.

BTW: I'd rather have a 32bit digital mixer (like in the M-Audio card) in a sound card than an analogue one.

For the sake of accuracy, M-Audio claims 36 (not 32) bit internal precision with the Audiophile 24/96's mixer.

Anyway... yeah, I suppose I would too. Still, there's something about keeping an analogue source fully analogue for listening purposes, without inserting two steps of conversion. I think I'll try a cheap A/B switch from Radio Shack (it's made for 2 inputs to 1 output, but should work for one input to 2 outputs unless they put diodes in there). Either that, or I'll build my own switchbox in a dual mono configuration, which might be slightly better in terms of crosstalk and potential ground loop issues.

this could be an awkward design, however isn't it possible that it is by default A -> A but because you chose a different sample rate, it switched to A -> D -> A to make sure that the sample rate chosen by you is enforced (it may not be well aware of the source, so it's being redundant)?

But what's a "different" sample rate on a semipro 24/96 soundcard? All the sample rates are just sample rates, afaik -- I don't think any resampling is taking place at <96KHz (someone correct me on that).

Anyway, changing other parameters like the card's internal clock for S/PDIF also affects the sound, so I don't think that's it.

contradiction? so what is it that's changing and made you think there is A->D->A converter in works?

(1) Switching between bitrates on the Hardware Settings tab produces audible clicks, and a temporary volume change
(2) Switching to rates below 44.1KHz is easily audible as the highest possible frequency changes (e.g. 16@32KHz or 11.025@22.05 KHz)
(3) Changing the output level between "Consumer" and -10dBV affects volume
(4) Changing the master clock to "S/PDIF In" kills the sound

None of the above would occur if the analog inputs were being routed directly to the analog outputs (as you might imagine when the Patchbay/Router tab says H/W Out 1/2 --> H/W In 1/2).

I believe that the signal goes through AD and then DA. I think that what you are modifying is the routing of the digital signals inside the card, not any other thing.

Yep, it seems so. I did end up buying an A/B switch, and am using this arrangement:

(1) Phono preamp goes to input "A" of the switchbox ("A"nalog)
(2) Soundcard output goes to input "B" of the switchbox
(3) Output of the switchbox goes to the headphone amp

(4) Main output of my preamp (also the phono stage) goes to the analog input of the soundcard for recording purposes.

After all this trouble -- I really can't tell the difference between the LP pumped through the soundcard, and the LP going directly to the headphone amp! The only real advantage is that I can listen to LP's with the computer turned off using this arrangement.

Who knows how much stereo crosstalk the cheapie ($13.95) A/B switch is adding, too... >_< I would have been safer building my own dual-mono switchbox, but... I didn't really want to do it.

1) Download RMAA 5.1 http://audio.rightmark.org/download.html

2) Test in loop mode & in loop mode while passing through the switch. You can test preamps or even amps this way if you build a line level converter (http://www.hut.fi/Misc/Electronics/circuit...er_to_line.html)

Now this is unbelievable... I opened the switchbox to take a look at the circuit board (it's Radio Shack Cat# 15-1952) and there were two resistors on the board! (in a switchbox?). After cutting them off, the sound went out in one channel on each A/B switch, so I just soldered jumpers across where they were -- everything is working fine.

Anyone have a clue why resistors would be put into an A/B switchbox? Maybe as a form of equipment protection if both switches are pressed at the same time, or something? I would think there should be no components, only circuit traces (or maybe a couple diodes under certain conditions). The two resistors were the only components in the box.

Anyway, I'm going to build a dual-mono switchbox eventually (have to find a drill, or some way to cut into a project box).